15-744: Computer Networking L-5 Congestion Control.
-
Upload
ilene-nash -
Category
Documents
-
view
218 -
download
0
Transcript of 15-744: Computer Networking L-5 Congestion Control.
L -4; 10-7-04© Srinivasan Seshan, 2004 2
News
• Note problem set 1 update• Part E: Please draw three RTTs after the loss,
not just one.
• Assigned reading• [JK88] Congestion Avoidance and Control• [CJ89] Analysis of the Increase and Decrease
Algorithms for Congestion Avoidance in Computer Networks
L -4; 10-7-04© Srinivasan Seshan, 2004 3
Congestion
• Different sources compete for resources inside network: Link b/w and queue space
• Why is it a problem?• Sources are unaware of current state of resource• Sources are unaware of each other• In many situations will result in < 1.5 Mbps of
throughput (congestion collapse)
10 Mbps
100 Mbps
1.5 Mbps
L -4; 10-7-04© Srinivasan Seshan, 2004 4
Causes & Costs of Congestion
• Four senders – multihop paths• Timeout/retransmit
Q: What happens as rate increases?
L -4; 10-7-04© Srinivasan Seshan, 2004 5
Causes & Costs of Congestion
• When packet dropped, any “upstream transmission capacity used for that packet was wasted!
L -4; 10-7-04© Srinivasan Seshan, 2004 6
Congestion Collapse
• Definition: Increase in network load results in decrease of useful work done
• Many possible causes• Spurious retransmissions of packets still in flight
• Classical congestion collapse• How can this happen with packet conservation• Solution: better timers and TCP congestion control
• Undelivered packets• Packets consume resources and are dropped elsewhere in
network• Solution: congestion control for ALL traffic
L -4; 10-7-04© Srinivasan Seshan, 2004 7
Other Congestion Collapse Causes
• Fragments• Mismatch of transmission and retransmission units• Solutions
• Make network drop all fragments of a packet (early packet discard in ATM)
• Do path MTU discovery
• Control traffic• Large percentage of traffic is for control
• Headers, routing messages, DNS, etc.
• Stale or unwanted packets• Packets that are delayed on long queues• “Push” data that is never used
L -4; 10-7-04© Srinivasan Seshan, 2004 8
Congestion Control and Avoidance
• Desirable properties:• Scalability:
• # flows, range of capacities, range of delays• Do well in entire range!
• Efficiency: High network utilization• Fairness• Works-ness: avoids collapse!
• Congestion collapse is not just a theory• Has been frequently observed in many
networks
L -4; 10-7-04© Srinivasan Seshan, 2004 9
Fairness
• Jain’s fairness index• f = (xi)2/n(xi
2)
• All x equal: 1• k/n get service: k/n
• Max-min fairness• No user receives more than their request, pi• No other allocation satisfying (1) has higher min allocation• condition 2 holds as we remove the minimal user & reduce
total resource accordingly• aka: ui = MIN(u fair, pi)
• Goal: Something that works well enough.
L -4; 10-7-04© Srinivasan Seshan, 2004 10
Objectives
• Simple router behavior • Distributedness
• Efficiency: Xknee = xi(t)
• Fairness: (xi)2/n(xi2)
• Power: (throughput/delay)• Convergence: control system must be
stable
L -4; 10-7-04© Srinivasan Seshan, 2004 11
Design questions
• Congestion• How is congestion signaled?
• Either mark or drop packets
• When is a router congested?• Drop tail queues – when queue is full• Average queue length – at some threshold
• Control questions:• How do senders react to congestion?• How do senders determine capacity for flow?
L -4; 10-7-04© Srinivasan Seshan, 2004 12
Congrol 2
• Upon congestion, flows must reduce rate• How? Decrease algorithm
• If no congestion, flows might try sending more. Increase algorithm.
• Let’s assume window-based flow control• sender maintains “cwnd”: # of unacknowledged
packets in the network at any time• Transmission rate: cwnd / rtt
• (Alternate: Rate-based; equation-based)
L -4; 10-7-04© Srinivasan Seshan, 2004 13
Linear Control
• Many different possibilities for reaction to congestion and probing• Examine simple linear controls• Window(t + 1) = a + b Window(t)• Different ai/bi for increase and ad/bd for
decrease• Supports various reaction to signals
• Increase/decrease additively• Increased/decrease multiplicatively• Which of the four combinations is optimal?
L -4; 10-7-04© Srinivasan Seshan, 2004 14
Phase plots
• Simple way to visualize behavior of competing connections over time
Efficiency Line
Fairness Line
User 1’s Allocation x1
User 2’s Allocation
x2
L -4; 10-7-04© Srinivasan Seshan, 2004 15
Phase plots
• What are desirable properties?• What if flows are not equal?
Efficiency Line
Fairness Line
User 1’s Allocation x1
User 2’s Allocation
x2Optimal point
Overload
Underutilization
L -4; 10-7-04© Srinivasan Seshan, 2004 16
Additive Increase/Decrease
T0
T1
Efficiency Line
Fairness Line
User 1’s Allocation x1
User 2’s Allocation
x2
• Both X1 and X2 increase/decrease by the same amount over time• Additive increase improves fairness and additive
decrease reduces fairness
L -4; 10-7-04© Srinivasan Seshan, 2004 17
Multiplicative Increase/Decrease
• Both X1 and X2 increase by the same factor over time• Extension from origin – constant fairness
T0
T1
Efficiency Line
Fairness Line
User 1’s Allocation x1
User 2’s Allocation
x2
L -4; 10-7-04© Srinivasan Seshan, 2004 18
What is the Right Choice?
• Constraints limit us to AIMD• Can have multiplicative term in increase (MAIMD)• AIMD moves towards optimal point
x0
x1
x2
Efficiency Line
Fairness Line
User 1’s Allocation x1
User 2’s Allocation
x2
L -4; 10-7-04© Srinivasan Seshan, 2004 19
TCP and linear controls
• Upon congestion:• w(t+1) = a*w(t) 0 < a < 1
• While probing• w(t+1) = w(t) + b 0 < b << wmax
• TCP sets a = 1/2, b = 1 (packet)
L -4; 10-7-04© Srinivasan Seshan, 2004 20
TCP Congestion Control
• Motivated by ARPANET congestion collapse• Underlying design principle: packet conservation
• At equilibrium, inject packet into network only when one is removed
• Basis for stability of physical systems
• Why was this not working?• Connection doesn’t reach equilibrium• Spurious retransmissions• Resource limitations prevent equilibrium
L -4; 10-7-04© Srinivasan Seshan, 2004 21
TCP Congestion Control - Solutions
• Reaching equilibrium• Slow start
• Eliminates spurious retransmissions• Accurate RTO estimation• Fast retransmit
• Adapting to resource availability• Congestion avoidance
L -4; 10-7-04© Srinivasan Seshan, 2004 22
TCP Congestion Control
• Changes to TCP motivated by ARPANET congestion collapse
• Basic principles• AIMD• Packet conservation• Reaching steady state quickly• ACK clocking
L -4; 10-7-04© Srinivasan Seshan, 2004 23
AIMD: Now you grok the sawtooth
• Distributed, fair and efficient• Packet loss is seen as sign of congestion and
results in a multiplicative rate decrease • Factor of 2
• TCP periodically probes for available bandwidth by increasing its rate
Time
Rate
L -4; 10-7-04© Srinivasan Seshan, 2004 24
Congestion Avoidance
• If loss occurs when cwnd = W• Network can handle 0.5W ~ W segments• Set cwnd to 0.5W (multiplicative decrease)
• Upon receiving ACK• Increase cwnd by (1 packet)/cwnd
• What is 1 packet? 1 MSS worth of bytes• After cwnd packets have passed by
approximately increase of 1 MSS
• Implements AIMD
L -4; 10-7-04© Srinivasan Seshan, 2004 25
Congestion Avoidance Sequence Plot
Time
Sequence No
Packets
Acks
L -4; 10-7-04© Srinivasan Seshan, 2004 26
Congestion Avoidance Behavior
Time
CongestionWindow
Packet loss+ Timeout
Grabbingback
Bandwidth
CutCongestion
Windowand Rate
L -4; 10-7-04© Srinivasan Seshan, 2004 27
Packet Conservation
• At equilibrium, inject packet into network only when one is removed• Sliding window and not rate controlled• But still need to avoid sending burst of packets
would overflow links• Need to carefully pace out packets (ack clocking)!• Helps provide stability
• Need to eliminate spurious retransmissions• Accurate RTO estimation• Better loss recovery techniques (e.g. fast retransmit)
L -4; 10-7-04© Srinivasan Seshan, 2004 28
TCP Packet Pacing
• Congestion window helps to “pace” the transmission of data packets
• In steady state, a packet is sent when an ack is received• Data transmission remains smooth, once it is smooth• Self-clocking behavior
Pr
Pb
ArAb
ReceiverSender
As
L -4; 10-7-04© Srinivasan Seshan, 2004 29
Reaching Steady State
• Doing AIMD is fine in steady state but slow…
• How does TCP know what is a good initial rate to start with?• Should work both for a CDPD (10s of Kbps or
less) and for supercomputer links (10 Gbps and growing)
• Quick initial phase to help get up to speed (slow start)
L -4; 10-7-04© Srinivasan Seshan, 2004 30
Slow Start Packet Pacing
• How do we get this clocking behavior to start?• Initialize cwnd = 1• Upon receipt of every
ack, cwnd = cwnd + 1• Implications
• Window actually increases to W in RTT * log2(W)
• Can overshoot window and cause packet loss
L -4; 10-7-04© Srinivasan Seshan, 2004 31
Slow Start Example
1
One RTT
One pkt time
0R
2
1R
3
4
2R
567
83R
91011
1213
1415
1
2 3
4 5 6 7
L -4; 10-7-04© Srinivasan Seshan, 2004 32
Slow Start Sequence Plot
Time
Sequence No
.
.
.
Packets
Acks
L -4; 10-7-04© Srinivasan Seshan, 2004 33
Return to Slow Start
• If packet is lost we lose our self clocking as well• Need to implement slow-start and congestion
avoidance together
• When timeout occurs set ssthresh to 0.5w• If cwnd < ssthresh, use slow start• Else use congestion avoidance
L -4; 10-7-04© Srinivasan Seshan, 2004 34
TCP Saw Tooth Behavior
Time
CongestionWindow
InitialSlowstart
Fast Retransmit
and Recovery
Slowstartto pacepackets
Timeoutsmay still
occur
L -4; 10-7-04© Srinivasan Seshan, 2004 35
TCP Modeling
• Given the congestion behavior of TCP can we predict what type of performance we should get?
• What are the important factors• Loss rate
• Affects how often window is reduced
• RTT• Affects increase rate and relates BW to window
• RTO• Affects performance during loss recovery
• MSS • Affects increase rate
L -4; 10-7-04© Srinivasan Seshan, 2004 36
Simple TCP Model
• Some additional assumptions• Fixed RTT• No delayed ACKs
• In steady state, TCP losses packet each time window reaches W packets• Window drops to W/2 packets• Each RTT window increases by 1 packetW/2
* RTT before next loss• BW = MSS * avg window/RTT = MSS * (W +
W/2)/(2 * RTT) = .75 * MSS * W / RTT
L -4; 10-7-04© Srinivasan Seshan, 2004 37
Simple Loss Model
• What was the loss rate?• Packets transferred = (.75 W/RTT) * (W/2 *
RTT) = 3W2/8• 1 packet lost loss rate = p = 8/3W2
• W = sqrt( 8 / (3 * loss rate))
• BW = .75 * MSS * W / RTT• BW = MSS / (RTT * sqrt (2/3p))
L -4; 10-7-04© Srinivasan Seshan, 2004 38
TCP Friendliness
• What does it mean to be TCP friendly?• TCP is not going away• Any new congestion control must compete with TCP
flows• Should not clobber TCP flows and grab bulk of link• Should also be able to hold its own, i.e. grab its fair share, or it
will never become popular
• How is this quantified/shown?• Has evolved into evaluating loss/throughput behavior• If it shows 1/sqrt(p) behavior it is ok• But is this really true?
L -4; 10-7-04© Srinivasan Seshan, 2004 39
TCP Performance
• Can TCP saturate a link?• Congestion control
• Increase utilization until… link becomes congested
• React by decreasing window by 50%• Window is proportional to rate * RTT
• Doesn’t this mean that the network oscillates between 50 and 100% utilization?• Average utilization = 75%??• No…this is *not* right!
L -4; 10-7-04© Srinivasan Seshan, 2004 40
Summary Unbuffered Link
t
W Minimum window for full utilization
• The router can’t fully utilize the link• If the window is too small, link is not full• If the link is full, next window increase causes drop• With no buffer it still achieves 75% utilization
L -4; 10-7-04© Srinivasan Seshan, 2004 41
TCP Performance
• In the real world, router queues play important role• Window is proportional to rate * RTT
• But, RTT changes as well the window
• Window to fill links = propagation RTT * bottleneck bandwidth
• If window is larger, packets sit in queue on bottleneck link
L -4; 10-7-04© Srinivasan Seshan, 2004 42
TCP Performance
• If we have a large router queue can get 100% utilization• But, router queues can cause large delays
• How big does the queue need to be?• Windows vary from W W/2
• Must make sure that link is always full• W/2 > RTT * BW• W = RTT * BW + Qsize• Therefore, Qsize > RTT * BW
• Ensures 100% utilization• Delay?
• Varies between RTT and 2 * RTT
L -4; 10-7-04© Srinivasan Seshan, 2004 43
Single TCP FlowRouter with large enough buffers for full link utilization
L -4; 10-7-04© Srinivasan Seshan, 2004 44
Important Lessons
• How does TCP implement AIMD?• Sliding window, slow start & ack clocking• How to maintain ack clocking during loss recovery
fast recovery
• Modern TCP loss recovery• Why are timeouts bad?• How to avoid them? fast retransmit, SACK
• How does TCP fully utilize a link?• Role of router buffers
L -4; 10-7-04© Srinivasan Seshan, 2004 46
Example
• 10Gb/s linecard• Requires 300Mbytes of buffering.• Read and write 40 byte packet every 32ns.
• Memory technologies• DRAM: require 4 devices, but too slow. • SRAM: require 80 devices, 1kW, $2000.
• Problem gets harder at 40Gb/s• Hence RLDRAM, FCRAM, etc.
L -4; 10-7-04© Srinivasan Seshan, 2004 47
Summary Buffered Link
t
W
Minimum window for full utilization
• With sufficient buffering we achieve full link utilization• The window is always above the critical threshold• Buffer absorbs changes in window size
• Buffer Size = Height of TCP Sawtooth• Minimum buffer size needed is 2T*C
• This is the origin of the rule-of-thumb
Buffer
L -4; 10-7-04© Srinivasan Seshan, 2004 48
Rule-of-thumb
• Rule-of-thumb makes sense for one flow• Typical backbone link has > 20,000 flows• Does the rule-of-thumb still hold?
L -4; 10-7-04© Srinivasan Seshan, 2004 49
If flows are synchronized
• Aggregate window has same dynamics• Therefore buffer occupancy has same dynamics• Rule-of-thumb still holds.
2maxW
t
max
2
W∑
maxW∑
maxW
L -4; 10-7-04© Srinivasan Seshan, 2004 50
If flows are not synchronized
ProbabilityDistribution
B
0
Buffer Size
∑W
L -4; 10-7-04© Srinivasan Seshan, 2004 51
Central Limit Theorem
• CLT tells us that the more variables (Congestion Windows of Flows) we have, the narrower the Gaussian (Fluctuation of sum of windows)
• Width of Gaussian decreases with
• Buffer size should also decreases with
n
CT
n
BB n ×
=→ = 21
n
1
n
1